• ArbitraryValue@sh.itjust.works
    link
    fedilink
    English
    arrow-up
    1
    arrow-down
    1
    ·
    edit-2
    2 days ago

    Art has some unusual properties, compared to other things people might want (homes, jewelry, etc). In economic terms, it’s a “non-rivalrous good” - it doesn’t get used up when people experience it. It’s also a good that is extremely easy to share with almost everyone in the world, and the world is a big place so there are a lot of very talented artists even though they’re only a small fraction of humanity.

    All together, this means that there’s not much reason to look at anything expect really good art - there’s more really good art available for free than you’ll have time to see, so why would you look at anything mediocre? The AI’s art is usually mediocre (although I’ve seen some pretty good pieces) so the people making a big deal of it look foolish. Don’t they realize that it’s just a poor imitation of human art?

    The thing is that the AI is better at art than almost every human is. It’s certainly faster and cheaper than any human. That’s not world-changing because of the unusual properties art has which I discussed, but it’s not world-changing in the same way that the Wright brothers’ first airplane wasn’t world-changing. It barely got off the ground and it certainly couldn’t do anything useful… Still, if birds could be jealous of flight then they should have been worried.

    So yeah, I think it’s amazing that a computer can create even a mediocre painting. Less than five years ago, it couldn’t. That’s already faster than most human artists learned to create art. I suspect that in another five years, its output will be indistinguishable from professional-quality human art even by experts.

    (And art is just a toy that AI companies use to show off their technology, like IBM with Deep Blue playing chess against Gary Kasparov. The real fun will be elsewhere.)